Decision Support for What-If Analysis and the Confirmation Bias

Abstract

The effects of the confirmation bias on users of a Decision Support System (DSS) for what-if analysis were explored through a laboratory experiment to determine how those users' decision processes were affected when they were predisposed toward a particular solution. A theoretical model was developed based on the hypotheses that users' predispositions affect their decision processes and their decisions and that users' decision processes affect their decisions. This model was tested using structural equation modeling techniques. Findings indicated that DSS users' predispositions did not directly affect the conclusions of the analyses they performed, but only their expectations of what those results would indicate to be the best solution. An additional finding was that DSS users' decisions are based directly on their recollections of the analyses they perform as opposed to the actual conclusions of the analyses. A modified model which should be of interest to researchers and practitioners is presented.